I didn’t phrase my argument very well, so let me try again. My argument is that a pure community development model based solely on volunteer labor doesn’t work very well.
You have to have some paid developers to maintain a good project, that is capable of attracting outside community developers. If you look at who actually maintains the Linux kernel, PHP, Apache, PostgreSQL, and most of the FOSS projects that work well, there are a number of developers who work at companies which pay them to spend part of their day working on these projects. Maybe they are working at a chip company or a company to creates web pages or a database services company, but part of their paid jobs is contributing to FOSS projects which are needed by the companies. When a FOSS project is well maintained and useful to people, it will attract more unpaid volunteers to work on it.
I think the mixture of paid and volunteer devs is what makes FOSS development work well. I don’t think that a pure community development model with nobody being paid works well, when trying to do core development of a new mobile desktop environment like Phosh. Ordinary volunteers may develop an app or two, but it is hard to attract quality volunteers to jump into the hard-core development of phoc, phosh and libhandy.
Ubuntu Touch isn’t really functioning right now, because it can’t attract these kinds of quality volunteers that can do development of the core system. It only exists because Canonical created it. There are only 9 users registered to make commits to the UBports’ project on Github. There have only been 29 commits to the Ubuntu Touch project since UBports took over in May 2017 and only only one of those happened in the last year:
Does that look sustainable? The same is true of LuneOS which has almost no commits, and basically just picked up the pieces of WebOS abandoned by HP. Maemo Leste, which also picked up the pieces of Hildon left by Nokia, only has 4 developers, according to a talk given by one of the devs last year. Glacier UI used by Nemo Mobile has been glacial in terms of its development speed.
The only mobile interface whose development is sustainable and is based on a community development model is KDE Plasma Mobile. It is mostly volunteer labor, but it is getting some corporate help from Blue Systems, plus Qt Group, SUSE and IBM/Red Hat contribute to the KDE Plasma libraries and some of the Kirigami apps that uses. With mostly volunteer labor, it is worth considering how far Plasma Mobile has advanced in 5 years, compared to how far Phosh has advanced in 2.5 years.
According to this poll, only 9% of PinePhone users are using Plasma Mobile, whereas 51% are using Phosh and 16% are using Ubuntu Touch. In other words, the two mobile interfaces that were developed by companies have almost all the users. (All the people who selected “Other” in the poll appear to either be using the command line or desktop interfaces, so they don’t count in terms of mobile interfaces.)
I don’t know how many new developers have been attracted to the Plasma Mobile project from outside KDE, but it is clear to me that Phosh has done better at attracting outside community developers than Ubuntu Touch. I count two devs from GNOME, two devs from postmarketOS and two devs from Mobian that now contribute to Phosh. When a project is advancing because it has paid developers doing the heavy lifting (but who are willing to work with the community), it tends to attract quality community volunteers who want to participate in a project that is making progress and making something useful.
Let’s ask where would the PinePhone be if it only relied only on a community development model. The only mobile interfaces it would have would be Plasma Mobile and Glacier IU, which currently only 9% are using. The PinePhone is viable, because it relies on the paid work done by Canonical in the past and Purism in the present.
Of course, Purism also needs the PinePhone, because it is bringing more users and developers to Phosh, and they are doing the work of porting Phosh to different distros, which is critical for the long-term success of Phosh. However, if PINE64 is going to play the role of attracting a larger user base to Phosh, then Purism shouldn’t play that role. Instead, Purism should play the role of paying for the core development, and in order to do that, Purism simply has to charge high prices for the Librem 5. If Purism stops paying developers, then we are in a situation like Plasma Mobile, where development won’t advance very fast.
My interest is getting mobile Linux to the point where it is useful for normal people. As I see it, we won’t get there if we wait for the community development of Plasma Mobile or if Purism abandons development of Phosh and hopes the community will take over. Our best shot at making mobile Linux a viable alternative to Android and iOS is if Purism is paying for the development of Phosh, but trying to build a larger community to contribute. In order for that to happen, some segment of people has to be willing to pay high prices for their Linux phones, because there aren’t hundreds of thousands of users to distribute the costs of paid software development. At some point, Purism may be able to sell 250k phones per year and charge a mid-range price for the Librem 5, but as long as Purism is selling 10k phones per year, it simply has to charge a high price if we expect Purism to be be able to pay for any serious dev work.
Currently Purism is only able to pay 7 software developers to work on the Librem 5, which isn’t enough in my opinion. To make Phosh into a mobile interface that can attract tens of thousands of new users to the platform and hundreds of new developers who will create apps, I think that Purism needs to be able to pay the salaries of at least 20 developers.
Remember that the goal is not to get a cheaper Linux phone next year, but to make mobile Linux into a viable alternative to Android and iOS and Phosh in my opinion has the best shot of getting us there of all the existing mobile interfaces. If we truly want mobile Linux to succeed in the long term and to expand software freedom and the right to privacy on mobile phones, we need to get Phosh to the point where its functionality is good enough to get people to switch from Android and iOS. We want the major phone makers to jump into the Linux phone market and adopt Phosh. To make all that happen, we need a lot of software development over the next couple years, and I think the fastest way to make that happen is for some customers to pay high prices for their Linux phones to cover the development costs.
For these reasons, I don’t think it is a good idea for us as buyers of the Librem 5 to demand that Purism aim for the mid-range with the Librem 5 v2–not if we are thinking strategically about the development of mobile Linux as a viable platform. I suspect that $799 is too high of a price on the demand curve to get the maximum profits to pay for the most software development, but microeconomics was never my strong suit, so I may be wrong. Still the principal remains that we should be thinking about what Purism needs to be able to pay for more developers.
